Reduction of Pulmonary Surfactant in Patients with Human Immunodeficiency Virus Infection and Pneumocystis carinii Pneumonia* Alan G. D. Hoffman, M.D., Ph.D.; Marion G. Lawrence, M.S.; Frederick P. Ognibene, M.D.; Anthony F. Suffredini, M.D.; Gregg Y. Lipschik, M.D.; Joseph A. Kovacs, M.D.; Henry Masur; M.D.; and james H. Shelhamer; M.D., F.C.C.P. We assessed qualitative and quantitative differences in surfactant lipid composition of bronchoalveolar lavage (BAL) 8uid in patients with acquired immune deficiency syndrome (AIDS) and Pneumocyatia corinii (PC) pneumonia. Five normal volunteers and 27 patients with human immunodeficiency virus (HIV) infection underwent BAL for evaluation of possible pulmonary infection. Bronchoalveolar lavage studies in eight patients were negative for PC organisms, and 19 were positive. Pneumocystia corinii pneumonia was graded (mild vs moderate to severe) by initial alveolar-arterial oxygen gradient. Bronchoalveolar lavage 8uid was centrifuged, the lipids were extracted from the supernatant, and total lipid profiles of dephosphorylated glycerolipids were analyzed as trimethylsilylether derivatives by high temperature gas-liquid chromatography. Phospholipase A. levels were determined using a radiolabeled
pneumocystis carinii (PC) pneumonia is the leading cause of serious lung disease in patients infected with the human immunodeficiency virus (HIV). 1 Although advances have been made in early diagnosis, 2 prophylaxis3 and therapy, •.~ this disease still has a mortality rate of 10-30 percent. 6 In patients with respiratory failure, and especially those who require mechanical ventilation, the mortality rate is 60 percent or greater. 7 Mortality due to PC pneumonia has been correlated with a variety of factors including elevation of serum lactate dehydrogenase, severity of chest radiograph abnormality, and an increased alveolararterial (A-a) oxygen gradient greater than 30 mm Hg. s-Io Pneumocystis carinii pneumonia severe enough to cause respiratory failure is most often characterized by marked deterioration in oxygenation due to intrapulmonary shunting of blood 11 and by a marked reduction in pulmonary compliance.U These detrimental physiologic changes are similar to abnormalities noted in the infant respiratory distress syndrome and in the adult respiratory distress syndrome, processes which may be related in part to a deficiency of *From the Critical Care Medicine Department, Warren G. Magnuson Clinical Center, National Institutes of Health, Bethesda. Manuscript received February 4; revision accepted AprillO. Reprint requests: Dr. Ho.ffrrwn, Critical Care Medicine Department, National Institutes of Health, Bldg 10 Rm 7043, Bethesda 20892
1730
E coli membrane method. Compared to the normal vol-
unteers (109± 13 p.g/5 ml) and the PC negative group (107 ± 13 p.g/5 ml), total BAL lipid was reduced for both the mild PC pneumonia group (73± 10 p.g/5 ml) and the moderate to severe PC pneumonia group (46±4 p.g/5 ml). There was a parallel reduction of diacylglycerol lipids: normal volunteers, 52± 7 p.g/5 ml; PC negative, 52± 9 p.gl 5 ml; mild PC pneumonia, 35 ± 7p.g/5 ml; and moderate to severe PC pneumonia, 15 ± 2 p.g/5 mi. Phospholipase A. activity in moderate to severe PC pneumonia was twice that of the PC negative patients, and 30 times that for normals. The data demonstrate a marked diminution in surfactant glycerophospholipid in patients with AIDS and PC pneumonia and suggest a potential role for surfactant abnormality in the pathophysiology of this disease. (Chest 1992; 102:1730-36)
pulmonary surfactant or of surfactant activity. 12• 15 Experimental studies in animal models of PC pneumonia have produced negative or conflicting data on changes in pulmonary surfactant and on the activation of phospholipases which might metabolize surfactant lipid. 16•17 To date, no studies are available on surfactant composition in patients with PC pneumonia. The purpose of this study was to examine bronchoalveolar lavage fiuid (BAL) from HIV positive patients with and without PC pneumonia to determine qualitative and quantitative changes in surfactant lipid composition and to assay for phospholipase ~ activity in BAL. METHODS
lhtient lbpu/ation
Twenty-seven adult patients from the Warren G. Magnuson Clinical Center, National Institutes of Health, and five normal volunteers were studied. The normal volunteers were negative for HIV risk factors and disease by history, and had normal chest radiographs and screening laboratory test results. All of the patients were infected with HIV as determined by EUSA (Dupont, Boston) and Western blot (Biotech, Rockville, MD) methods using standard kits. The 32 subjects underwent diagnostic bronchoscopy either for the evaluation of suspected pulmonary disease or as part of an approved Clinical Center protocol.
Bronchoscopy Flexible fiberoptic bronchoscopy was performed in a standardized Reduction ol Pulmonary Surfactant In Patients with HIV and PCP (Hoffman et el)
manner as previously described.'" All patients received intravenous sedation as well as topical anesthesia with 1 percent lidocaine. After the airways were examined, one segment was lavaged with 180 ml of0.9 percent sodium chloride in six equal aliquots. Fluid recovery averaged 50 percent of the total volume instilled. Samples were sent to the microbiology and cytology laboratories for stains and culture. The diagnosis of PC pneumonia was based on visualization of the organisms using toluidine blue 0 or monoclonal antibodybased indirect immunofluorescent stains of a centrifuged pellet from BAL, 1 or toluidine blue 0 or methenamine silver stains of transbronchial biopsies. Patients were stratified into five groups based on the results of the bronchoscopy and A-a oxygenation gradient: (I) normal volunteers; (2) HIV + , asymptomatic, no Pneumocystis; (3) HIV +, symptomatic, no Pneumocystis; (4) mild PC pneumonia (A-a 0 1 gradient 30 mm Hg).
Studies of BAL Upid After centrifugation (1,500 X G, 10 min to remove cells), an aliquot of BAL supernatant fluid (5 ml) was lyophilized. Tridecanoylglycerol (50 mg) was added as an internal standard. Prior to running on the gas chromatograph, the diacyl phospholipids and monoacyl (lyso) phospholipids were digested to their respective diacylglycerols and monoacylglycerols by incubation with phospholipase C from Clostridium welchii (Sigma Chemical Co, St Louis). The lipids were extracted with chloroform-methanol (2:1) by the
method of Folch et al.'" The lower phase was dried by trituration through anhydrous sodium su1fate, and the solvent removed at 40"C under nitrogen. Lipids were converted to their trimethylsiyl (I'MS) ethers using TriSyl BSA (Pierce Chemical Co, Rockford, Ill)-' before injection into the gas chromatograph. Lipids were resolved by class and carbon number using hightemperature gas-liquid chromatography, as described by Kubis et al,• on a Varian 3700 series gas chromatograph (Varian Instruments, Sunnyvale, Cal) equipped with a flame-ionization detector, dual difFerential electrometer. The samples (1-5 ml) were injected directly onto 2ft x 'AI inch OD stainless-steel columns packed with 3~ OV-1 on 100'120 mesh Gas-Chrom Q (Supelco, Bellefonte, Pa). The injector and detector were maintained at 320"C and 360"C respectively, while the column was temperature programmed in a linear manner at 4°C/min from ISO"C to 350"C, with initial and final times of 4 min each. Helium was used as the carrier gas at a flow of 40 mllmin. Peaks were identified and quantified by relative retention times using the internal standard method, following calibration of the instrument with a standard mixture of neutral lipids including acylglycerols, sterols, steryl esters and fatty acids.
Phoapholipa.fe A. Assay Phospholipase A. activity was assayed by the radiolabeled membrane technique of Patriciarca et aJu as modified by Vadas. • E coli (strain 011) was grown in brain-heart infusion media containing 1-['"C]-oleic acid (New England Nuclear, Boston). Thin-layer chro-
Table 1-Charactm.tict of32lbtienla Undergoing Bronclao.copg Category Group 1 Normal volunteers
Group2 HIV+IPCneg Asymptomatic Group3 HIV+IPC neg Symptomatic
Group4 HIV+IPCP Mild
GroupS HIV+IPCP Mod/Severe
Patient No.
HIV Status
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
NT NT NT
22 23 24 25
26 27 28 29 30 31 32
Symptoms
NT NT + + + + + + + + + + + + + + + + + + + + + + + + + + +
+ + + + + + + + + + + + + + + + + + + + + + + +
A-a Gradient mmHg* 6 0 0 17 t t t 20 21 10 45 36 t 10 10 8 10 19 6 23 27 33 33 34 38 38 43 43
54 62
68 74
p carinii
Other Brooch Results Normal BAL NormalBAL NormalBAL NormalBAL Normal BAL Normal lung tissue Interstitial pneumonitis Interstitial pneumonitis Interstitial pneumonitis Interstitial pneumonitis Interstitial pneumonitis Interstitial pneumonitis Interstitial pneumonitis
+ + + + + + + + + + + + + + + + + + +
*Initial A-a gradient on presentation before therapy. tHemoglobin 98-1~ saturated by pulse oximetry. NT= Not tested, normal volunteers negative by history, chest x-ray exam and screening lab tests. CHEST 1102 I 6 I DECEMBER, 1992
1731
matography and enzymatic digestion confinned that this tracer was preferentially incorporated into the sn-2 position of E coli glycerophospholipids. The assay mixture included: E coli suspension {250 p.l = 100,000 DPM}, 20 percent bovine serum albumin {250 p.l), 0.1 M Tris-HCI buffer containing 10 mM CaCI. (pH 8.5, 450 p.l) and human BAL supernatant {100 p.l). Incubation was carried out for 15 min at 37"C in a Buchler vortex mixer {Buchler Instruments, Fairfield, NJ), and the reaction finally stopped by cooling the samples to 4°C. The mixture was then filtered through a 0.22 mm pore-size membrane to separate the albumin-bound released free fatty acids from the undigested 1-["C)-oleic acid labeled membrane bound phospholipids. Filtrate counts were quantified by liquid scintillation counting using 10 ml Aquasol 2 (Dupont, Boston). After subtraction of appropriate background and blank controls, data were converted to arbitrary units of enzyme activity where 1 unit = release of 25 percent of initial E coli labeV15 min. Statistical Analyses
Data are expressed as means ± standard error of the mean. Comparisons between sample groups were made using either Student's t-test for unpaired samples"" or standard analysis of variance (ANOVA) followed by the a posteriori Student-NewmanKeuls {SNK) test" for comparison of experimental group means. Rejection of the null hypothesis was made with probabilities of less than 5 percent {p value
pneumocystis carinii (PC) pneumonia is the leading cause of serious lung disease in patients infected with the human immunodeficiency virus (HIV). 1 Although advances have been made in early diagnosis, 2 prophylaxis3 and therapy, •.~ this disease still has a mortality rate of 10-30 percent. 6 In patients with respiratory failure, and especially those who require mechanical ventilation, the mortality rate is 60 percent or greater. 7 Mortality due to PC pneumonia has been correlated with a variety of factors including elevation of serum lactate dehydrogenase, severity of chest radiograph abnormality, and an increased alveolararterial (A-a) oxygen gradient greater than 30 mm Hg. s-Io Pneumocystis carinii pneumonia severe enough to cause respiratory failure is most often characterized by marked deterioration in oxygenation due to intrapulmonary shunting of blood 11 and by a marked reduction in pulmonary compliance.U These detrimental physiologic changes are similar to abnormalities noted in the infant respiratory distress syndrome and in the adult respiratory distress syndrome, processes which may be related in part to a deficiency of *From the Critical Care Medicine Department, Warren G. Magnuson Clinical Center, National Institutes of Health, Bethesda. Manuscript received February 4; revision accepted AprillO. Reprint requests: Dr. Ho.ffrrwn, Critical Care Medicine Department, National Institutes of Health, Bldg 10 Rm 7043, Bethesda 20892
1730
E coli membrane method. Compared to the normal vol-
unteers (109± 13 p.g/5 ml) and the PC negative group (107 ± 13 p.g/5 ml), total BAL lipid was reduced for both the mild PC pneumonia group (73± 10 p.g/5 ml) and the moderate to severe PC pneumonia group (46±4 p.g/5 ml). There was a parallel reduction of diacylglycerol lipids: normal volunteers, 52± 7 p.g/5 ml; PC negative, 52± 9 p.gl 5 ml; mild PC pneumonia, 35 ± 7p.g/5 ml; and moderate to severe PC pneumonia, 15 ± 2 p.g/5 mi. Phospholipase A. activity in moderate to severe PC pneumonia was twice that of the PC negative patients, and 30 times that for normals. The data demonstrate a marked diminution in surfactant glycerophospholipid in patients with AIDS and PC pneumonia and suggest a potential role for surfactant abnormality in the pathophysiology of this disease. (Chest 1992; 102:1730-36)
pulmonary surfactant or of surfactant activity. 12• 15 Experimental studies in animal models of PC pneumonia have produced negative or conflicting data on changes in pulmonary surfactant and on the activation of phospholipases which might metabolize surfactant lipid. 16•17 To date, no studies are available on surfactant composition in patients with PC pneumonia. The purpose of this study was to examine bronchoalveolar lavage fiuid (BAL) from HIV positive patients with and without PC pneumonia to determine qualitative and quantitative changes in surfactant lipid composition and to assay for phospholipase ~ activity in BAL. METHODS
lhtient lbpu/ation
Twenty-seven adult patients from the Warren G. Magnuson Clinical Center, National Institutes of Health, and five normal volunteers were studied. The normal volunteers were negative for HIV risk factors and disease by history, and had normal chest radiographs and screening laboratory test results. All of the patients were infected with HIV as determined by EUSA (Dupont, Boston) and Western blot (Biotech, Rockville, MD) methods using standard kits. The 32 subjects underwent diagnostic bronchoscopy either for the evaluation of suspected pulmonary disease or as part of an approved Clinical Center protocol.
Bronchoscopy Flexible fiberoptic bronchoscopy was performed in a standardized Reduction ol Pulmonary Surfactant In Patients with HIV and PCP (Hoffman et el)
manner as previously described.'" All patients received intravenous sedation as well as topical anesthesia with 1 percent lidocaine. After the airways were examined, one segment was lavaged with 180 ml of0.9 percent sodium chloride in six equal aliquots. Fluid recovery averaged 50 percent of the total volume instilled. Samples were sent to the microbiology and cytology laboratories for stains and culture. The diagnosis of PC pneumonia was based on visualization of the organisms using toluidine blue 0 or monoclonal antibodybased indirect immunofluorescent stains of a centrifuged pellet from BAL, 1 or toluidine blue 0 or methenamine silver stains of transbronchial biopsies. Patients were stratified into five groups based on the results of the bronchoscopy and A-a oxygenation gradient: (I) normal volunteers; (2) HIV + , asymptomatic, no Pneumocystis; (3) HIV +, symptomatic, no Pneumocystis; (4) mild PC pneumonia (A-a 0 1 gradient 30 mm Hg).
Studies of BAL Upid After centrifugation (1,500 X G, 10 min to remove cells), an aliquot of BAL supernatant fluid (5 ml) was lyophilized. Tridecanoylglycerol (50 mg) was added as an internal standard. Prior to running on the gas chromatograph, the diacyl phospholipids and monoacyl (lyso) phospholipids were digested to their respective diacylglycerols and monoacylglycerols by incubation with phospholipase C from Clostridium welchii (Sigma Chemical Co, St Louis). The lipids were extracted with chloroform-methanol (2:1) by the
method of Folch et al.'" The lower phase was dried by trituration through anhydrous sodium su1fate, and the solvent removed at 40"C under nitrogen. Lipids were converted to their trimethylsiyl (I'MS) ethers using TriSyl BSA (Pierce Chemical Co, Rockford, Ill)-' before injection into the gas chromatograph. Lipids were resolved by class and carbon number using hightemperature gas-liquid chromatography, as described by Kubis et al,• on a Varian 3700 series gas chromatograph (Varian Instruments, Sunnyvale, Cal) equipped with a flame-ionization detector, dual difFerential electrometer. The samples (1-5 ml) were injected directly onto 2ft x 'AI inch OD stainless-steel columns packed with 3~ OV-1 on 100'120 mesh Gas-Chrom Q (Supelco, Bellefonte, Pa). The injector and detector were maintained at 320"C and 360"C respectively, while the column was temperature programmed in a linear manner at 4°C/min from ISO"C to 350"C, with initial and final times of 4 min each. Helium was used as the carrier gas at a flow of 40 mllmin. Peaks were identified and quantified by relative retention times using the internal standard method, following calibration of the instrument with a standard mixture of neutral lipids including acylglycerols, sterols, steryl esters and fatty acids.
Phoapholipa.fe A. Assay Phospholipase A. activity was assayed by the radiolabeled membrane technique of Patriciarca et aJu as modified by Vadas. • E coli (strain 011) was grown in brain-heart infusion media containing 1-['"C]-oleic acid (New England Nuclear, Boston). Thin-layer chro-
Table 1-Charactm.tict of32lbtienla Undergoing Bronclao.copg Category Group 1 Normal volunteers
Group2 HIV+IPCneg Asymptomatic Group3 HIV+IPC neg Symptomatic
Group4 HIV+IPCP Mild
GroupS HIV+IPCP Mod/Severe
Patient No.
HIV Status
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
NT NT NT
22 23 24 25
26 27 28 29 30 31 32
Symptoms
NT NT + + + + + + + + + + + + + + + + + + + + + + + + + + +
+ + + + + + + + + + + + + + + + + + + + + + + +
A-a Gradient mmHg* 6 0 0 17 t t t 20 21 10 45 36 t 10 10 8 10 19 6 23 27 33 33 34 38 38 43 43
54 62
68 74
p carinii
Other Brooch Results Normal BAL NormalBAL NormalBAL NormalBAL Normal BAL Normal lung tissue Interstitial pneumonitis Interstitial pneumonitis Interstitial pneumonitis Interstitial pneumonitis Interstitial pneumonitis Interstitial pneumonitis Interstitial pneumonitis
+ + + + + + + + + + + + + + + + + + +
*Initial A-a gradient on presentation before therapy. tHemoglobin 98-1~ saturated by pulse oximetry. NT= Not tested, normal volunteers negative by history, chest x-ray exam and screening lab tests. CHEST 1102 I 6 I DECEMBER, 1992
1731
matography and enzymatic digestion confinned that this tracer was preferentially incorporated into the sn-2 position of E coli glycerophospholipids. The assay mixture included: E coli suspension {250 p.l = 100,000 DPM}, 20 percent bovine serum albumin {250 p.l), 0.1 M Tris-HCI buffer containing 10 mM CaCI. (pH 8.5, 450 p.l) and human BAL supernatant {100 p.l). Incubation was carried out for 15 min at 37"C in a Buchler vortex mixer {Buchler Instruments, Fairfield, NJ), and the reaction finally stopped by cooling the samples to 4°C. The mixture was then filtered through a 0.22 mm pore-size membrane to separate the albumin-bound released free fatty acids from the undigested 1-["C)-oleic acid labeled membrane bound phospholipids. Filtrate counts were quantified by liquid scintillation counting using 10 ml Aquasol 2 (Dupont, Boston). After subtraction of appropriate background and blank controls, data were converted to arbitrary units of enzyme activity where 1 unit = release of 25 percent of initial E coli labeV15 min. Statistical Analyses
Data are expressed as means ± standard error of the mean. Comparisons between sample groups were made using either Student's t-test for unpaired samples"" or standard analysis of variance (ANOVA) followed by the a posteriori Student-NewmanKeuls {SNK) test" for comparison of experimental group means. Rejection of the null hypothesis was made with probabilities of less than 5 percent {p value
